This work is to evaluate the potential use of solid-phase microextraction (SPME) in the analysis of tobacco components, mainly alkaloids. In SPME, sampling can be done either by taking vapor phase in the headspace of a tobacco sample or by dipping the SPME fiber into an aqueous tobacco extract. Because of the low volatility of alkaloids, sampling from headspace of a dry or a damped tobacco sample requires additional effort to improve extraction efficiency. Thus, extraction of tobacco samples with an aqueous solution is also tested. A SPME fiber coated with polydimethylsiloxane was directly dipped into the tobacco extract, then injected onto a capillary column under split or splitless mode (with 1 or 2 mm i.d. injection liners). When combined with the practice of fast GC, nicotine and a group of selected minor alkaloids ( i.e. , nornicotine, myosmine, anabasine and anatabine) can be separated with baseline resolution within 3 min. In addition, the carry-over problem frequently occurring in alkaloids analysis was eliminated. However, certain limitations in SPME applications need to be addressed. For example; SPME is an equilibrium process and its extraction efficiency is highly dependent upon the partitioning of target compounds between the polymeric phase and the sample matrix. The matrix effect can be compensated by the use of an internal standard, by the addition of individual target compound, or by the use of isotope-labeled internal standards as reported in the literature. In addition, the solvent-free injection of SPME essentially eliminates the refocusing effect of solvent on peak broadening, particularly for those early eluted peaks. Peak broadening and resolution in SPME analysis are also influenced by the mass transfer of target compounds, which is determined by temperature, pH, the agitation of solution, and film thickness of polymer coating on the fiber. The influence of the above factors and the optimization of experimental parameters will be discussed.